Folks,

I am wondering that i can use something like this in one line.
For example, $((cat filename > wc -l) / 2)
It doesn't work; how to get it work using command substitution?

Moreover, is there any option for wc -l not to return filename after the line counts?
wc -l filename would return, for instance, 10 filename

I want the line counts only.

Thanks a lot

You can use sed to strip away the file name. For example, this will sum up the line count for an entire directory and save it as $I.

I=0

for LINES in $(wc -l * | sed -e "s/^ *\([0-9]*\) .*/\1/")
do
(( I += LINES ))
done

echo $I

The tr and cut commands could also be used as in

... for LINE in $(wc -l * | tr -s " " "\t" | cut -f2) ...

But for very large files, the line count could extend to the left-most column and you'd end up cutting the wrong field.

Your example would be written as:

expr $(wc -l filename | sed -e "s/^ *\([0-9]*\) .*$/\1/") / 2

wc -l < filename

(( count = $(wc -l < filename) / 2 ))

A far more elegant solution by reborg. Quite nice.